CZ196696A3 - Apparatus for suppressing optical distortion caused by four-wave interaction - Google Patents

Abstract

An optical telecommunication system comprises at least two sources of optical signals modulated at different wavelengths, having respective coherence times; a multiplexer for multiplexing said signals in a single common optical fibre; an optical-fibre line connected at one end to said common optical fibre of said multiplexer; means for receiving said signals; an element for reducing the four wave mixing (FWM) between said signals, optically connected in series along said optical-fibre line, wherein said FWM-reducing element comprises an optical circuit including at least two-wavelength-selective filters, each relative to a wavelength band including one of said optical signals, said filters being optically connected in series with each other over an optical path, the length of at least one section of said optical path included between two consecutive filters being greater than the length corresponding to the coherence time of at least one of said sources of optical signals. <IMAGE>

Description

BACKGROUND OF THE INVENTION The present invention relates to an optical circuit for suppressing optical distortion resulting from a four-wave interaction between optical signals along an optical communication line, an optical amplifier comprising such a four-wave interaction suppression circuit, an optical communication system with reduced level of distortion. transmission of optical signals with reduced level of distortion caused by four-wave interaction.

BACKGROUND OF THE INVENTION

A four-wave interaction, also known as a four-photon interaction or four-wave mixing (FWM), is a non-linear third-order phenomenon causing the generation of a new signal arising from the interaction between three existing signals. The frequency fp of the newly generated signal is bound to the frequencies fj_, fj, f _{to the} interacting signals by the equation:

f = fi, f, f _k.

Maximum efficiency of generating a new signal or fourth wave corresponds to a situation where the polarization of interacting signals is the same and the condition of phase synchronism is met

Δβ = P (f) + p (fj) - 3 (f _k) P (f _F) = o where β (f) is the propagation constant of the signal at frequency f.

The interacting signals are not necessarily three separate signals. A four-wave interaction can also occur as a result of interaction between signals from two sources (degenerate case). In the presence of only two interacting signals at frequencies e.g. fy _and f _k , signals at frequencies f _F = 2fi-f _k and _F = 2f _k -fj may _be generated.

Four-wave interaction is an obstacle to fiber optic communication using the wave multiplex (WDM) method. According to this method, several mutually independent communication channels, each associated with a specific wavelength, are transmitted simultaneously by an optical fiber transmission line.

Due to the above-mentioned non-linear third-order phenomenon and under high radiation intensity conditions in the fiber core (especially amplification) and considerable interaction lengths between signals, signal generation as a result of intermodulation between signal pairs or triplets may occur when FWM is used. Such a phenomenon is described, for example, in Journal of Lightwave Technology, Vol. 8, No. 9, September 1990, pages 1402-1408.

The wavelengths of the generated signals may lie in the band used for communication channels, especially when they may be the same or lie close to the wavelength of one of the channels; the probability of this phenomenon increases sharply with the increasing number of communication channels used.

It is known that the light signals emitted by the optical fiber line are attenuated as they pass, which necessitates amplification by means of corresponding amplifiers located at predetermined intervals along the line.

For this purpose, optical amplifiers are preferably used in which the signal remains in optical form and is amplified and does not need to be detected and recreated. Said optical amplifiers are based on the properties of fluorescent dopants such as erbium, which, when suitably excited by light pumping energy, exhibit high emission in the wavelength band, where quartz-based optical fibers exhibit minimal attenuation.

In the case of a multistage cascade of optical amplifiers by a four-wave interaction in each communication line s, the signals generated by the stages are amplified in the same way as the communication signals and added to the signals generated by the four-wave interaction at the other stages. At the end of the line, the signals generated by the four-wave interaction at each of these stages are summed: if the FWM signals are phase-close, then the total FWM signals coming from the sum can reach such intensity that prevents proper reception of the communication signals.

The optical fibers used along the communication line exhibit a chromatic dispersion due to a combination of properties related to the refractive index profile and the material forming said fibers, where the dispersion varies with changes in the wavelength of the transmitted signal and becomes zero for a given value of λθ of said wavelength.

This phenomenon of chromatic dispersion is mainly manifested by an increase in the duration of the signal generating pulses that propagate along the fiber, which is due to different propagation rates of the different chromatic components of each pulse, characterized by their own wavelength propagating along the fiber at different speeds.

This elongation causes time-successive pulses well differentiated at the time of radiation to occur. after passing through the fiber on the receiving side, partially overlap so that they are no longer recognizable as separate events and cause a reception error.

So-called dispersed dispersed fibers (DS) are known whose optical properties are chosen such that the zero chromatic dispersion point wavelength lies between 1500 and 1600 nm generally used in telecommunications.

Fibers of this type are defined in the ITU-T Recommendation

G.653 of March 1993, wherein the chromatic dispersion in the fiber is zero nominally for a wavelength Xg of 1550 nm with a tolerance of nm to the above value.

DS fibers are described, for example, in US Patent Nos. 4,715,679, 4,822,399, 4,755,022, and are commercially available from CORNING Inc., Corning, NY (US) under the trade names SMF / DS (registered trademark) and FIBER OTTICHE SUD SpA, Battipaglia (IT) under the trade name SM DS.

In particular, it should be noted that the synchronβ - 0 phase synchronism condition defined above is met and the FWM-generated signals reach a high intensity if the wavelength of one of the communication signals is equal to or very close to the wavelength Xg of zero fiber dispersion. are distributed symmetrically around Xg.

A method designed to address the problem of FWM intermodulation between signals in multichannel systems, described in IEEE Photonics Technology Letters,

Vol.3, No.6, June 1991, pages 560-563 is based on the use of communication signals with different polarization. This method is very complex because it is necessary to adjust the polarization of each signal with respect to the communication line; its efficiency is further limited by the fact that generally used optical fibers do not maintain the plane of polarization during transmission.

FC4, published in OFC / IOOC '93 Technical Digest, pages 252-253, proposes the distribution of optical frequencies of individual communication channels at different distances with uneven distribution; these frequencies are selected such that the signals generated by the four-wave interaction between possible pairs or triplets of the communication signals have frequencies far enough from the communication signals and can be separated from them by a filter. However, this method significantly reduces the usefulness of the bandwidth (wavelength) usable for communication purposes (compared to the case of equidistantly distributed channels); moreover, a high signal wavelength stability is required which requires the use of an additional device to control said stability.

Unequal wavelength distribution of FWM suppression channels in an optical WDM communication system is also suggested in PRMorkel US Patent No. 5,410,624 in combination with spectral regeneration means of optical signals comprising an optical circulator and a string of narrow-reflection fiber grids, each tuned to one of

Ί multiplexed wavelengths, wherein said string is connected to an intermediate gate of an optical circulator.

The third method disclosed in Electronic Letters, Vol.30, No.11, 26/05/95, pages 876-878, is made for use on a communication line consisting of fiber sections of low absolute dispersion value having alternately positive and negative value. When applying this method, we cannot use existing communication lines and new lines need to be created. Moreover, the assembly of the necessary new lines is difficult, since we have to choose among the already produced fibers those having dispersion values suitable for use in different sections and to take into account the difficulty of choosing between already produced fibers with constant dispersion and the required accuracy limits.

Another method is provided in an article by K. Inoue published in the Journal of Lightwave Technology in March 1993, Vol.11, No.3, pages 455-461. It is based on creating a random relationship between the phases of the generated FWM signals along the various stages of the multichannel communication line so as to prevent incremental channel counting.

To create a random phase relationship, an optical circuit is provided that is placed immediately before or after each amplifier along a communication line made up of M fiber optic sections and (Μ-1) amplifiers.

The optical circuit is composed of a demultiplexer, which splits the communication channels according to frequency, optical paths of different length and multiplexer, where the optical signals are combined into a single output. The optical paths connected between the outputs of the demultiplexer and the inputs of the multiplexer are selected so that the difference between the lengths of any two is greater than the coherence length I _c = ν / π zdrojeν of the source signal, where v is the light speed in the medium. The signals generated in each of the stages by the four-wave interaction are uncorrelated to each other and therefore add together at the end of the line in power rather than in amplitude as found in conventional systems. Under the condition of phase synchronism (Δβ = Ο), the total radiation power generated by the FWM and generated along the fiber optic transmission line is lower compared to the case without the optical circuit in a ratio corresponding to the number of optical fiber sections in the line.

The aforementioned article mentions that the described method can be used in multichannel optical communication systems in which individual communication channels are selected from the transmission line by demultiplexing channels of different wavelengths and over different optical paths, connecting the channel to be selected to the receiver, and the remaining channels they are brought back to the common path after multiplexing.

The paper shows the possibility of immediate application of the described method in multichannel communication systems equipped with demultiplexers capable of selecting individual channels from a line, while other channels continue their journey through a common optical path. An example of this type of demultiplexer is

introduced Fabry - Perot circulator. filter combined with optical The applicant could however observe that optical circuit suggested in the article for suppression of four waves interaction and

in particular, the demultiplexers required to separate signals at different wavelengths are difficult to perform for practical assembly, especially when using a large number of channels. A possible solution by deploying several demultiplexers in a cascade with fewer outputs would create more complex and voluminous devices with different attenuations in different channels.

The large volume could also be caused by the large overall length of the line sections connecting the multiplexer and demultiplexer, especially in the case of a large number of communication channels. In fact, the fiber required for channel n must be at least n times longer than the coherence length of the radiation source and the sum of the lengths of these sections at least N (N + 1) / 2 times the coherence length, where N is the total number of communication channels.

In addition, a circuit constructed according to the instructions of the aforementioned article cannot be modified to a new configuration if one or more channels need to be added or changed to a wavelength of one or more channels; in this case it is necessary to replace the demultiplexer and multiplexer.

U.S. Patent No. 5,283,686 to D.R. Hubera introduces, inter alia, a WDM optical communication system comprising an optical amplifier, an optical circulator and optical fiber filters with a Bragg grating, one for each communication channel. The system allows suppression of spontaneous emission at wavelengths other than the communication channel. The patent does not mention the problem of distortion created by the four-wave interaction along the communication line.

Patent application MI94A0002556 filed on December 16, 1994 on behalf of the same applicant relates inter alia to an optical communication system comprising:

at least two optical signal sources modulated at different wavelengths in a predetermined transmission bandwidth and at a predetermined transmission rate;

means for multiplexing said signals to enter a single optical fiber;

an optical fiber coupled at one end to the multiplexing means;

means for receiving said signals, comprising optical demultiplexing means for the signals themselves depending on the wavelength;

wherein said signals have an optical power value greater than a predetermined value in at least one portion of said optical fiber conduit comprising an optical fiber having a chromatic dispersion lower than a predetermined value in said wavelength transmission band, characterized in that said optical fiber has a chromatic dispersion increasing with increasing wavelength and showing a zero value at a wavelength lower than the minimum wavelength of said band by such a value that no local wavelength with zero dispersion in the fiber and the possibility of generating products due to the four-wave interaction lies within said band.

The problem arises in the practical implementation of an optical device capable of reducing the power of radiation induced by the four-wave interaction between signals propagating along the communication line and in which the use of optical fibers other than conventional DS fibers along the communication line is not required.

SUMMARY OF THE INVENTION

In one aspect, the present invention relates to an optical communication system comprising:

at least two optical signal sources modulated at different wavelengths and having corresponding coherence times;

a multiplexer for multiplexing said signals into a single common optical fiber;

- an optical fiber line connected at one end to said common optical fiber of said multiplexer;

- means for receiving said signals associated with a second end of said optical fiber line and comprising a demultiplexer of said optical signals;

an element for suppressing the four-wave interaction between said signals, optically connected in series along said optical fiber line;

wherein said FWM suppression element comprises an optical circuit comprising at least two wave-selective filters, each corresponding to a wavelength band comprising one of said optical signals, said filters being optically connected to each other in series in an optical path, the length of at least one section of said optical path, which is interposed between the two following filters is greater than a length corresponding to a coherence time of at least one of said optical signal sources.

In particular, said optical circuit for each of said optical signals comprising a wave selective filter related to a wavelength band that includes a corresponding optical signal and does not include the remaining optical signals, said filters are optically coupled to each other in series in an optical path; the two subsequent filters being greater than a length corresponding to the coherence time of each of said optical signal sources.

In particular, said optical circuit comprises an optical circulator having an input / output gate connected to said optical fiber line and at least one input / output gate connected to one of said wave selective filters.

In particular, said wave-selective filters are Bragg lattice filters especially made of optical fiber.

In one assembly positioned along said optical fiber there is at least one optical amplifier, which preferably comprises one active optical fiber doped with a fluorescent dopant, in particular an erbium, and a pumping radiation source.

Said FWM suppression element may preferably be optically connected in series at an intermediate position along said active optical fiber.

In this case, the optical path for pumping radiation that is supplied from outside said FWM suppression element may be provided between two sections into which said active optical fiber is divided by said element.

Alternatively, said optical amplifier may comprise two active optical fiber sections, each with its own pumping radiation source.

In particular, the lengths of said active optical fiber sections, the fluorescent dopant concentration in the active optical fiber, and the pumping power sources are operatively selected such that the total gain of said amplifier differs by less than 2dB from the same amplifier without said FWM suppression element. continuity between the two active optical fiber sections has been restored.

In a second aspect, the present invention relates to a process for transmitting optical signals comprising:

generating two modulated optical signals with corresponding wavelengths;

multiplexing said signals at one end of an optical transmission line comprising at least one single-mode optical fiber section in which intermodulation occurs by a four-wave interaction between the signals;

receiving said signals at the other end of said optical transmission line;

and comprising a selective delay operation of said signals at intermediate positions along said optical transmission line, wherein said selective delay operation comprises:

selectively transmitting said signals to corresponding optical paths of predetermined lengths, the lengths of which are adapted to phase decorrelate said signals to one another;

combining said signals after passing said optical paths;

- in which at least one section of said optical path is common.

In particular, said process comprises the step of optically amplifying said signals at least once along said transmission line. Preferably, said step of selectively delaying said signals is immediately preceded and followed by the steps of optically amplifying said signals.

In a third aspect, the present invention relates to an optical amplifier comprising:

a first and a second active optical fiber doped with a fluorescent dopant;

pumping means for said first and second active optical fibers adapted to deliver pumping power;

- coupling means between said first optical active fiber for coupling said optical pumping power and at least two transmission signals of different wavelengths and corresponding coherence times,

- a four-wave interaction suppression element between said signals, said element being optically connected in series between said first and second active optical fibers and said FWM suppression element comprising an optical circuit comprising optical paths of different lengths to which said broadcast signals are selectively transmitted, and said lengths are such that at least two of the signals undergo a relative delay greater than the corresponding coherence times.

Preferably, at least one section of said optical paths is common to said signals.

Said fluorescent dopant is especially erbium and preferably said active fiber comprises aluminum, germanium and lanthanum as additional dopants.

BRIEF DESCRIPTION OF THE DRAWINGS

More details can be seen from the following description with reference to the attached drawings, where:

Fig. 1 is a diagram showing an optical communication system with a wavelength multiplex according to the invention;

Fig. 2 is a diagram showing a two-stage optical line amplifier;

Fig. 3 is a diagram of an optical circuit according to the invention for suppressing optical distortion caused by FWM;

Fig. 4 is a diagram of an experiment with a device object of the present invention;

Fig. 5 is a power spectral curve passing through two Bragg grating optical fiber filters used in the experimental apparatus of Fig. 4 and connected in series;

6A, 6B are graphs obtained in the experiment and numerical simulation with respect to the standardized power of the FWM signals generated in the presence of two communication signals when the wavelength of one of the signals in the first (A) and second (B) stages of the experimental apparatus of FIG. .4;

Fig. 7 is a graph showing the normalized power of the FWM signal measured during the aforementioned experiment when changing the wavelength of one of the transmitted channels with and without the distortion suppression circuit and comparing it with a numerical simulation of the communication line case without the distortion suppression circuit;

Fig. 8 shows a diagram of a two stage pump line amplifier according to the invention for suppressing optical distortion caused by FWM;

Fig. 9 is a schematic of a single pumping stage line amplifier according to the invention including an optical circuit for suppressing optical distortion caused by FWM;

Fig. 10 is a diagram of an optical circuit according to the invention for suppressing optical distortion caused by FWM along a bidirectional optical communication line.

Examples

The wavelength multiplex optical telecommunications system will now be described with reference to FIG.

This description is related to the case of a telecommunications system using four independent communication channels with different wavelengths. However, this particular case is used only as an example: the following description is valid unless otherwise specified, for the general case of any number of communication channels of different wavelengths.

The communication system of FIG. comprises a transmitting station 3 formed by optical signal sources 1, 1 ', 2, 2', each having a different wavelength λμ λ2, λ £ ', λ2' ', located in the usable working range of the amplifiers, which are subsequently located in the system , and the line width Δνμ ÁV2, Δν2 ', ÁV2' '·

The optical signals are supplied to a signal combiner 81 adapted to simultaneously transmit signals at wavelengths λμ λ2, a single optical fiber input.

Generally, the signal combiner 81 is a passive optical device in which the optical signals transmitted by the corresponding optical fibers are summed into a single fiber; devices of this type are, for example, assembled from bonded fiber couplers, planar optics, microoptics, etc. available on the market.

Through the fiber 82, the optical signals are sent to a booster 83 where the signal level is increased to a level sufficient for transmission by the adjacent adjacent fiber optic section up to the next amplification medium and maintaining at the end a power level sufficient to provide the desired transmission quality.

Further, a booster 83 is connected to a fiber optic section 84a typically comprised of a step change refractive index and introduced into a suitable type of fiber optic cable of several tens (or hundreds) kilometers, for example, about 100 km of length.

Connected to the end of said first optical fiber section 84a is a first optical circuit 10a described below, which is adapted to suppress optical distortion caused by intermodulation between communication channels due to a four-wave interaction. The output of the optical circuit 10a is terminated in a first line amplifier 85a which is adapted to receive the Sigrals attenuated as they pass through the fiber and amplify them to a level sufficient to power the second optical fiber section 84b having the same characteristics as the preceding section.

The following optical distortion suppression circuits 10b, 10c, line amplifiers 85b, 85c, 85d and optical fiber sections 84c, 84d, 84e cover the desired total transmission distance up to the receiving station 6 comprising a preamplifier 87 adapted to receive signals and amplify them to a power level suitable in terms of receiver sensitivity and loss compensation in the following demultiplexer.

From the preamplifier 87 are signals of the demultiplexer 88, wherein said signals are dependent on their wavelengths and then the receivers 89, 90, 90 ', 90' '.

aired split aired to v

to

The demultiplexer 88 is a device adapted to be distributed to several output fibers of optical signals supplied by a single input fiber and split them according to different wavelengths; the demultiplexer may consist of a fused fiber splitter that splits the input signal into signals in several output fibers, each of which passes through a corresponding bandpass filter located at the wavelength in question.

E.g. a component similar to the signal combiner 81 described above can be used when it is built in an inverted configuration and coupled to corresponding bandpass filters.

Band-pass filters of this type are, for example, commercially available from MICRON-OPTICS, INC., 2801 Buford Hwy, Suite 140, Atlanta, Georgia, US, the suitable model being FFP-100.

The described configuration provides particularly satisfactory results for transmissions over distances of the order of about 500 km at high transmission rates of, for example, 2.5 Gbit / s (achieved here with four multiplexed wavelengths equivalent to 10 Gbit / s over a single wavelength) using four line amplifiers, one booster and one preamplifier.

For the purposes of this invention and the above use, the booster 83 is, for example, a commercially available optical fiber amplifier having the following characteristics:

input power -13.5 to 3.5 dBm output power 12 to 14 dBm operating wavelength 1534 to 1560 nm

Booster does not contain a notch filter.

A suitable model is TPA / E-MW available from the applicant.

Said booster uses erbium-doped active optical fiber Al / Ge / Er type.

As a booster, an amplifier operating under saturation conditions is proposed wherein the output power depends on the pumping power as described in detail in European patent EP 439,867, which is incorporated herein by reference.

For the purposes of the present invention and the above use, preamplifier means an amplifier located at the end of a line capable of amplifying the signal supplied to the receiver to a value sufficiently above the sensitivity threshold of the receiver itself (e.g. from -26 to -11 dBm at the receiver input) possible noise and maintain signal equalization.

E.g. the preamplifier 87 may be either a line amplifier using the same active fiber as the line amplifiers

85a-85d described below or a preamplifier specially designed for this purpose based on special requirements.

A suitable model is the RPA / E-MW supplied by the applicant.

The arrangement of the transmission system described above is particularly suitable for the desired function, in particular for the transmission of several WDM channels, if a particular selection of line amplifiers is made, especially in conjunction with the possibility of transmitting selected wavelengths without disadvantaging some of them over others.

The same behavior in all channels in the wavelength band between 1530 and 1560 nm is particularly guaranteed in the case of cascade amplifiers, when line amplifiers with particularly uniform (flat) characteristics at several different wavelengths are used in the cascade.

For the above purpose, the amplifier used as the line amplifier is formed according to the diagram of Fig. 2 and comprises one erbium doped active fiber 62 and a corresponding pumping laser 64 coupled to the dichroic coupler 63; the optical insulator 61 is positioned downstream of the fiber 62 in the propagation direction of the amplified signal, while the second optical insulator 65 is positioned downstream of the active optical fiber itself.

Said amplifier further comprises a second erbium-doped active fiber 66 and coupled to a corresponding pumping laser by a dichroic coupler 67; another optical insulator 69 is positioned toward the fiber 66.

An alternative not shown is a line amplifier in the form of a single stage amplifier according to the requirements of the respective application.

A preferred arrangement in which the above-described type of active fiber is used as a line amplifier, as described in detail in Italian Patent Application No.MI994A000712 of April 14, 1994 to the same Applicant, which is incorporated herein by reference, the contents of which are summarized below.

The composition and optical properties of active line amplifier fibers are summarized in the following table:

A1 ₂ O3 Ge02 La2 ° 3 Er ₂ O ₃ ON X _c wt% (mol%) wt% (mol%) wt% (mol%) wt% (mol%) nm 4 (2.6) 18 (11.4) 1 (0.2) 0.2 (0.03) 0.219 911

where:

wt% = (mean) percent oxide content in the core mol% = (mean) molar oxide content in the core

NA = numerical aperture (n] - 2 - n ₂ 2) 1/2

X _c = LP11 cut off.

Composition analyzes were performed in preform (before fiber drawing) by scanning electron microscope (SEM Hitachi).

Analyzes were performed at 1300x magnification at discrete points located along the diameter and spaced apart

200 μπι. Said fibers were produced by a vacuum deposition method in a quartz glass tube.

In said fibers, the incorporation of germanium dopant into the S1O2 matrix in the fiber core is achieved during synthesis.

The incorporation of erbium, aluminum and lanthanum into the fiber core is accomplished by a solution doping method wherein the synthesized fiber core material is embedded in a solution of dopant chloride in water prior to formation of the preform.

A more detailed description of the solution doping method is given, for example, in US Patent No. 5,282,079, which is incorporated by reference.

The pumping lasers 64, 68 are preferably quantum well lasers having the following characteristics: emission wavelength λρ = 980 nm maximum optical output power P _u = 80 mW Lasers of the above type are manufactured, for example, by LASERTRON Inc., 37 North Avenue, Burlington, MA (USA).

Dichroic couplers 63, 6Ί are fused-fiber couplers made from single-mode fibers at a wavelength of 980 nm and having an output power variation of less than 0,2 dB within the 1530 to 1560 nm band depending on polarization.

Dichroic couplers of the above type are known, commercially available and manufactured, e.g., by GOULD Inc, Fiber Optic Division, Baymeadow Drive, Gelm Burnie, DM (US) or SIFAM Ltd., Fiber Optic Division, Woodland Road,

Devon (GB).

The optical isolator 61, 65, 69 are optical insulators of a type that do not depend on the polarization of the transmitted signal with isolation greater than 35 dB and reflection losses below -50 dB.

Used insulators such as MDL 1-15 PIPT-AS / N 1016 supplied by ISOWAVE, 64 Harding Avenue, Dover, NNJ (US) or PIFI 1550 IP02 supplied by ETEK DYNAMICS Inc., 1885 Lundy Ave., San Jose, CA (US ).

The described line amplifier is capable of delivering a maximum total output power (signals plus spontaneous emission) of about 14 dBm at low signal gain of about 30 dB.

Under these operating conditions the total input power of the second stage is about 10 dBm and the second stage operates under saturation conditions. The total output power fluctuates by less than 0.2 dBm per dBm of the second stage total input power variation.

An apparatus for suppressing optical distortion caused by the four-wave interaction of the present invention is described below with reference to Fig. 3.

The optical circuit 10 comprises an optical circulator 15 with three access gates 11, 12, 13, a selective reflection wavelength optical filter 16 coupled to the optical circulator gate 12, and a selective delay circuit at a wavelength λ ₂ comprising a single-mode fiber section 18, one end of which is connected to the outlet of the filter 16 and the optical reflection filter 19 with selective reflection at the wavelength λ ₂ connected to the other end of the fiber section 18. The length of the single-mode optical fiber section 18 is equal to or greater than half the maximum value of I _c of the coherence lengths I / O, I / O ₂ , I / O ₂ ', I / O ₂ '', sources 1, 2, 2' Where v is the rate of propagation of radiation in the fiber.

A selective reflection filter at the wavelength λ of one of the communication signals in a WDM communication system means an optical component capable of reflecting a significant portion of the radiation at a wavelength within a predetermined wavelength and capable of transmitting a significant portion of the radiation outside said predetermined band. it comprises said wavelength λ and does not include wavelengths of the other transmitted signals.

The optical circuit 10 further comprises selective delay circuits 17 ', 17' ', each of which is comprised of a section 18',

18 '' single mode optical fiber, equal to or longer than 1 _c / 2, having one end connected to the filter outlet of the preceding selective delay circuit and the other end coupled to the optical filter 19 ', 14, with selective reflection at wavelength λ] /, kty '

The gates 11 and 13 of the optical circulator 15 are intended to be connected along an optical fiber communication line designed to transmit optical signals at wavelengths λχ, λ2, · ^ 2 ''

In particular, the circuit 10 is intended to be connected along a communication line with a cascade of optical amplifiers, described as an example of FIG. In this case, the gates 11 and 13 of the optical circulator 15 will be connected to the output of one of the optical fibers 84a-84d and the input of one of the line amplifiers 85a-85d.

The output of the last selective reflection filter (furthest from the optical circulator) shall be suitably terminated to prevent reflection of interfering radiation into the optical circulator. For this purpose, any of the conventional methods known to those skilled in the art, such as a low-reflection angled connector 24 may be used. A suitable connector is, for example, the FC / APC model manufactured by SEIKOH GIKEN, 296-1 Matsuhidai, Matsudo, Chiba (JP). If the optical circuit 10 is located along a communication line in a cascade with optical amplifiers, the termination may be formed by a dichroic coupler adapted to separate the residual pumping radiation of the amplifier connected in the direction of the circuit 10 from the spontaneous emission of the same amplifier, both can be monitored.

The optical connections between the various components of the optical circuit can be made by any of the known methods, e.g., buttwelding. Optical connections between different selective delay circuits 17, 17 ', 17' 'can also be made using optical connectors, preferably of low reflection type, so that removal or addition of other selective delay circuits can be easily accomplished.

Alternatively, it is possible to provide a selective delay line comprising an optical filter 16 and selective delay circuits 17, 17 ', 17''along a single optical fiber section by means of filters 16, 19, 19', 19 '' positioned along said optical fiber section in the optical fiber. spacings greater than I _c / 2 by the methods described below. The optical fiber section is therefore connected to the optical circulator gate 12. This alternative solution has the advantage of not requiring optical links between the different components of the selective delay line, so that corresponding attenuations are eliminated.

This selective delay line may also be provided in a modular version according to another alternative solution by assembling a plurality of selective reflection filters along optical fiber sections according to the methods described below at distances greater than 1 _c / 2. Optical fiber sections of this type may be assembled for various combinations of wavelengths and filter bands and may either alternatively be coupled to the optical connector portions 12 or interconnected in series depending on the number and characteristics of the channels to be used in the communication system.

In each of the described versions, the order of positioning of the selective reflection filters 16, 19, 19 ', 19' 'along the selective delay line is not critical to the present invention: this order may be changed during the assembly of said line.

Optical circulators are passive optical components, generally with three or four gates spaced in an ordered sequence, unidirectionally transmitting input radiation from each gate to the next successive gateways, and in particular to the adjacent successive gateway. The circulators used are preferably of a type whose response does not depend on polarization. Optical circulators are commercially available components. For example, the CR1500 model manufactured by JDS FITEL Inc, 570 Heston Drive, Nepean, Ontario (CA) or the PIFC-100 model manufactured by E-TEK DYNAMICS (already mentioned) are useful herein.

Selective reflection filters adapted for use in the present invention are, for example, Bragg grating waveguide filters. They reflect radiation in a narrow wavelength band and transmit radiation outside this band. They consist of a section of an optical waveguide, such as an optical fiber, along which the refractive index exhibits periodic changes: if parts of the signal reflect each change in refractive index in phase with others, constructive interference occurs and the input signal is reflected. The fulfillment of the condition of constructive interference corresponding to the maximum reflection can be expressed by 21 = λ ₃ / η, where 1 is the lattice constant of the lattice created by changes in the refractive index, λ ₃ is the wavelength of input radiation and n is the refractive index of the optical waveguide core. The described phenomenon is mentioned in the literature in connection with the Bragg lattice.

Periodic refractive index changes can be generated by known methods such as exposing the deprotected optical fiber portions with interference strips formed by an intense ultraviolet beam (such as an excimer laser, a frequency doubled argon laser, or a frequency quadruple Nd: YAG laser) that is capable of interfering with itself in a suitable interferometric system eg with a silicon phase mask as described in US Patent No. 5,351,321. The fiber, and in particular the core, is exposed to UV radiation with a periodically variable intensity along the optical axis. In the sections of the core affected by UV radiation, the Ge-0 bonds are partially disrupted, causing a permanent change in the refractive index.

By selecting the lattice constant to satisfy the constructive interference condition, the mean wavelength of the reflected band is determined as needed.

With this method it is possible to produce filters with -3dB wavelength band of 0.2 to 0.3 nm with reflection in the middle of the band up to 99%, the mean wavelength of the reflected band is definable during production with accuracy of + - 0.1 nm and varies with temperature only about 0.02 nm / degree.

The wavelengths of the sources 1, 2, 2 ', 2' 'have a tolerance band wider than + - 0.2 to 0.3 nm and bandwidth filters of corresponding width can be supplied. For sources based on, for example, commonly used semiconductor lasers, the emitted wavelength is determined with an accuracy of typically + - 1 nm.

Optical fiber filters with Bragg grating can be produced with the following characteristics: the reflection bandwidth can be greater than 0.2 to 0.3 nm when a variable grating constant grating is used according to a method known, for example, from PCHilla and others published in Electronic Letters, vol.30, No.14, from 7.7.1994, pages 1172-1174.

If it is necessary to compensate the chromatic dispersion at the wavelengths of the communication signals due to the operating conditions in the optical communication line in which the device according to Fig. 3 is used, then optical fiber filters with variable grating constant having Bragg grating F. Ouellette, published in Optics Letters, vol. 12, No.10, pages 847-849 of October 1987, can be used as selective reflection filters 16, 19, 19 ', 19' '.

If we use optical circuit 10 in an environment with significant temperature changes, fiber optic filters 16, 19,

19 ', 19' 'may require the use of stabilization.

The operation of the device in Fig. 3 can be described as follows: signals of wavelengths λ, λ, Xg ', β 2 enter through the gate 11 of the optical circulator 15 and propagate therefrom to the gate of the circulator 12. Then the signals enter the filter with selective Reflection 16. The wavelength signal is reflected from the circulator gate 12 and propagates to the circulator gate 13. The wavelengths lying outside the narrow band centered on λ] (all remaining wavelength signals λ2, λ2 ', λ2''are included in this radiation) instead pass through the filter 16 and pass through the single-mode optical fiber section 18 into the filter with a selective reflection 19 that reflects radiation at wavelength / 2 ^and transmits other signals. The wavelength signal is returned through the fiber and filter 16 to the optical circulator gate 12 where it is mixed with the wavelength signal. Due to the double passage of the optical fiber section 18, the wavelength signal λ 2 is phase decorated with respect to the wavelength signal λχ, and is phase-shifted by a length I _c greater than the coherence length of the sources 2. '' pass through the filter 19 and the optical fiber section 18 'to a filter 19' on which a signal of wavelength 1 2 is reflected.

Then this signal returns to the optical circulator gate 12 phase shifted by a length greater than 2 I _c relative to the signal at wavelength λχ and by a length greater than I _c relative to the signal at wavelength λ 2 so that it is phase shifted by a length greater than is the coherence length with respect to each of the two signals and is therefore phase decorated with both.

Finally, the signal at the wavelength Ίνχ ^{e is} reflected by a 19 '' filter. Due to the double passage of the optical fiber portion 18 ' ^{, it} is mixed with the other signals phase shifted by a length greater than the coherence length with respect to each of them and is therefore phase decorated with them.

Finally, the signals propagate from the optical circulator gate 12 to the circulator gate 13, which is connected to the optical communication line, as shown in Fig. 1.

The above description may be immediately extended to a plurality of signals of other wavelengths. The device can be easily adapted to function with as many communication signals at different wavelengths as needed by connecting one selective delay circuit for each communication channel following the first in any order.

This device can be easily modified after it has been manufactured to be re-assembled for the desired number of channels.

The above description of the optical distortion suppression device of FIG. 3 relates to a preferred embodiment of the same, wherein said device comprises one selective reflection filter for each transmitted communication signal.

In addition to this preferred arrangement, the device may be fabricated according to an alternative arrangement in which only some of the communication channels are separated from one-mode fiber lengths longer than half the coherence length of the signals to the selective reflection filters, while the remaining communication channels are reflected by one or more reflectors with sufficient reflection bandwidth. This configuration makes it possible to adjust some of the FWM signals generated towards the optical distortion reducing device to be phase decorated relative to the corresponding FWM signals generated away from the device.

Experiment

To verify the operation of the optical distortion suppression device, the Applicant compared the optical communication line according to the invention with the classical type of communication line during the experiment.

The experimental setup will now be described with reference to FIG.

Two coherent optical sources are labeled 31 and 32. They are two tunable semiconductor lasers with external resonator, model HP81678A manufactured by HEWLETT PACKARD Co., Rockwell, DM (US) and model TSL-80 manufactured by SANTEC, Micom Valley Tohkadai, Kamsue, Komaki Aichi 485 (JP). The line width of these sources is about Δν = 100 kHz. From the equation I _c = v / πΔν, where v is the velocity of radiation along the fiber, we obtain the value of I _c = 650 m coherence length in the fiber for the source e used.

The polarization of the signal from the source 32 was set equal to the polarization of the signal from the source 31 by means of a polarization control device located at the output of the source 32 and composed of two elements rotatable about an axis and formed by two coils of single-mode optical fiber 20-40 mm in diameter. The two signals with the same polarization are summed by the -3 Db coupler 34 and supplied to the booster 35 of the TPA / E-MW model previously mentioned and supplied by the applicant.

The amplified signals are transmitted to the first stage of the transmission line consisting of the dispersed dispersed optical fiber with a length of 13.8 km. The fiber has a zero dispersion around the wavelength of 1543 nm, the slope of the dispersion curve is about 0.1 ps / (nm ^ km) and the absorption is about 0.21 dB / km for the wavelengths used.

The optical distortion suppression circuit 20 has been coupled to the end of the optical fiber 36. It comprises an optical circulator 15, model CR1500, supplied by the aforementioned JDS FITE, connected by a gate 11 to the optical fiber 36; an optical fiber filter with a Bragg grating 16 connected to the optical circulator gate 12 and having a maximum reflection at a wavelength of 1543.7 nm; a single-mode optical fiber 18, which is connected to the outlet of the filter 16 and is 460 m long and is therefore greater than half the coherence length (about 325 m in the case of the sources used); an optical fiber filter with a Bragg lattice 19 attached to a fiber 18 with suitably selected spectral properties that were adjusted to the wavelengths of the signals during the experiment.

Figure 5 shows the spectral dependence 47 of the power reflected by the described filter 16 coupled in series with one of the filters 19 used during the experiment. For this measurement, the filters were connected to the central gate of the optical circulator according to the experimental arrangement in Fig. 4. The dependence 47, relative to the scale on the left side of the graph, shows the power measured at the output gate of said optical circulator while the spectral power signal shown in Fig. 5. Dependence 48 (relative to the scale on the left side of the graph) was present at the entrance gate of the circulator itself.

The optical circuit 20 further comprises a tunable interference filter 21 coupled to the output of the filter 19 and configured to transmit the wavelength of the generated FWM signal and to suppress components at the wavelengths of the residual signals after radiation of filters 16 and 19. The filter 21 was connected to input -3 dB coupler 22, the second input of which was connected to the gate 13 of the optical circulator 15. The connections between the optical elements were made by butt-welding.

The radiation generated by the FWM in the first stage passes through the filter 21 and coupler 22, which are not captured in the general circuit diagram of the distortion suppression circuit 10 described in Figure 3, to the second stage of the communication line.

The passage of this radiation is allowed for the following reason: the experiment is related to the case of communication signals at only two wavelengths; under these conditions, the signal generated by the FWM has a wavelength different from both signals and its propagation along the communication line can be suppressed by means of filters; this method of filtering FWM radiation could not be used, as previously shown, in the more general case of a plurality of communication channels where the wavelengths of the signals generated by the FWMs can collide with the wavelengths of the channels themselves and therefore cannot be filtered out; in order to carry out an experiment with only two communication channels to verify how the signals generated by the FWM are summed (this experiment should therefore be relevant for more general cases), it is necessary to ensure the passage of radiation generated by the FWM in the first stage to the second stage; the filter 21 and the coupler 22 allow this.

The -2dB couplers 22 and 34 are of the fused / f ibre type supplied by GOULD.

The following circuit 20 was a line amplifier 37, more particularly the Applicant's OLA / E-MW model based on erbium-doped active fiber optically pumped at a wavelength λ _ρ = 980 nm. This line amplifier has a gain of around 30 dB at a total input channel power of -20 dBm.

The total optical output power (signals plus amplified spontaneous emission), on the other hand, is about 12 to 14 dBm.

The signals coming from the line amplifier 37 were fed to a second stage transmission line consisting of a single-mode dispersed dispersed optical fiber 38 with a length Lg = 5.1 km. This fiber has a zero dispersion at a wavelength of 1545 nm, a slope of the dispersion curve is about 0.1 ps / (nm ² km) and an absorption at the wavelengths used about 0.21 dB / km.

After passing through the optical fiber 38, the signals were analyzed by an optical spectral analyzer 39, model MS9030A / MS9701B manufactured by ANRITSU Corp., 5-100-27 Minatoku, Tokyo (JP).

Giant. Figures 6A and 6B show power graphs of FWM signals generated along optical fibers 36 and 38 measured at a fixed wavelength of the source 31 as a function of the change in the wavelength of the source 32. In this measurement, each of the two fibers was connected directly to an adjustable attenuator connected to the amplifier 35 and a spectral analyzer 39 temporarily disabling the other components of the optical circuit and at a power of 2-4 mW in each channel fed to the fiber input. The power values in μΝ are shown in Figures 6A and 6B and standardized to input signals of 1 mW (0 dBm) power per channel according to:

PFWM ^ ^norm ^ PFWM / (P ⁱⁿ 31 · P ⁱⁿ 32) ² 'where pi ⁿ 3i, pi ⁿ 32 are the power values of the optical input carriers. The polarization of the signal from the source 32 was rotated by the device 33 to the FWM signal measured at the fiber output 36 or the fiber output 38 reached maximum.

For fiber 36 (FIG. 0A), the wavelength λλ of the optical source 31 was fixed at 1533.58 nm and the measurement of the power generated by the FWM was performed as a function of the wavelength change γ.

an optical source 32 between 1542.80 nm and 1543.80 nm in steps of approximately 0.05 nm. The measurement results are indicated by solid squares in Fig. A connected by line 41. The three peaks are clearly visible at 1543.05 nm, 1543.51 nm and 1543.66 nm; in the applicant's view, this phenomenon can be explained by the fact that at the above-mentioned wavelengths the dispersion value in the fiber 36 used in the experiment is zero and the phase synchronism condition is met.

The numerical simulation on the model described in Journal of Lightwave Technology, vol.10, No.11, November 1992, pages 1553-1561 was performed for an optical fiber 3 6 consisting of three segments of zero dispersion DS fiber at λ _ο 1 = 1543,05 nm , λ _ο 2 = 1543.51 nm, λ _ο 3 = 1543.66 nm.

The other parameters used in the three-segment fiber model are as follows:

attenuation α = 0,21 dB / km glass refractive index mode field average nonlinear 3rd order n = 1,45

MFD = 8 µm

X _im = 4.2 6.10 " ¹⁴ m ³ / J slope of the chromatic dispersion curve

Dc '= 0.1 ps / (nm2. Km)

The calculated normalized power of the intermodulation peaks Pj? V? 4 at the wavelengths of the signal λ = 1533, 58 nm and λ ₂ lying between 1542.80 nm and 1543.80 nm is shown in Fig. 6B of the dependencies 42.

Comparison of dependencies 41 and 42 shows that the FWM profile of intermodulation peaks expressed from the model is qualitatively similar to the experimentally obtained profile. From this it can be deduced that the optical fiber with the dispersion properties envisaged in the model used in the simulation represents an accurate model of the actual optical fiber 36 in the wavelength examined, particularly with respect to the generation of FWM signals.

Fig. 6B shows the corresponding results for fiber 38 used in the second stage of the experimental apparatus. The curve 43, connecting the experimentally measured points, captures the normalized power of the generated FWM signals originating from the intermodulation between a signal of wavelength λ ₂ = 1534.84 nm and a signal of wavelength λ ₂ varying from 1544.10 nm to 1546.00 nm in steps 0.1 nm. The two peaks are at 1544.80 nm and

1545.40 nm.

Also in this case, the experimental dependencies 43 and 44 coincide well with the numerical simulation of the optical fiber model 38, composed of two zero dispersion segments at wavelengths λθΐ - 1544, 80 nm and Xq2 = 1545, 40 nm and with other parameters identical to those which were used for optical fiber simulation 36.

The data obtained for the dispersion of the optical fibers 36 and 38 was used for subsequent numerical simulation of the overall behavior of the optical distortion suppression device shown in FIG. The simulated data were compared with the experimental results in Fig. 7.

Fig. 7 shows the standardized power p of FWM signals generated by intermodulation between the first signal of wavelength λχ = 1533.7 nm and the second signal of wavelength λ 2, which is variable. The results obtained are shown both with and without circuit 20 of FIG. 4 for suppressing optical distortion.

Experiments were performed for three values of the wavelength Ag of the second signal, λ2χ = 1545.50 nm, λ22 ⁼ 1546.70 nm and λ23 = 1547.80 nm. Instead of the filter 19 shown in the diagram of FIG. 4, the previously described butt-welding type having the center of the reflection band at the corresponding wavelength is used.

Numerical simulations were performed for Xg values between 1544 nm and 1549 nm.

The x-axis of the graph in Fig. 7 shows the value of Xg ·

On the other hand, the values on the y-axis are the values of the parameter p, which correspond to the standardized value of the difference between the power Pp (counter) (total PpwM power) of FWM signals measured at the device output and the sum of PfWM power (D ^af FWM (2); represents the power measured at the output of the second stage, the FWM signals generated separately along the first and second stages, i.e. along the fibers 36 and 38 according to the diagram in FIG.

More specifically, the parameter p is given by the following expression:

(iot) - 0) + (2))

2 · yjPpwM (0- Pfwm)

In order to estimate ρ, it was necessary to measure, in addition to the total power coming from the device for each experimental wavelength X2, the power of FWM signals coming from the second stage, which would be generated along the first and second stages without interference.

The optical power of the PWM ( ¹ ) ^FWM signal along the first stage exiting the device can be directly measured by temporarily connecting the spectrum analyzer 39 to the amplifier output or the second stage output and dividing this value by the known attenuation value of the fiber 38 constituting the second stage.

The optical power of the Pp (2) signal originating from the FWM along the second stage can be measured with the optical distortion suppression device 20 connected by temporarily disconnecting the optical link between the filter output 21 and the coupler 22 to prevent the FWM signals generated in the first stage to the second stage without affecting the channel power entering the second stage.

If the distortion suppressor circuit 20 is not connected, the optical power PjwM (2) is instead measured by temporarily replacing the first stage optical fiber 36 with an attenuator with adequate attenuation to suppress the generation of FWM signals without changing the channel power entering the second stage.

Points 51, 52, 53 in the graph of Fig. 7 show the experimental results at the three wavelengths λ2ΐ, λ22 ^and 2323 given the case where the optical distortion suppression circuit 20 is connected between the first and second stages of the device of Fig. 4.

It should be noted that the parameter p is permanently zero.

The total power generated by the FWM in the second stage corresponds for each wavelength λη to the sum of the FWM signals generated in both stages. Therefore, any additional phenomenon caused by interference between the two FWM signals at each stage is eliminated.

If the distortion suppression circuit 20 is not wired, the fiber output 36 is directly connected to the amplifier 37 input in the device of FIG. 4, but the above interference is effective, as confirmed by the measurements shown in the graph of FIG. 56. In both first cases, the total power of the FWM signals at the device output is greater than the sum of the power of the FWM signals generated in the first and second stages by about 80%.

In the case of measurement 56, the total power of the FWM signal at the device output relative to the wavelength of the second signal λ2 = λ23 is less than the sum of the power of the FWM signals in both stages; in this case, the interference of the two FWM signals from both stages is partially destructive and the connection of the distortion suppression optical circuit 20 increases the FWM distortion value than when the circuit 20 is disconnected.

This contradiction is only possible if it is generally not possible to accurately determine the signal wavelengths and dispersion of the optical transmission line. As generally explained, the best possible conditions cannot be assumed; on the contrary, it is always possible that, due to uncertainty or due to small changes in one of the parameters, a condition for positive interference between FWM signals in the individual stages appears.

The present invention avoids creating an unfavorable condition by limiting the power of the overall FWM signal relative to the sum of the power generated in each amplification stage.

The curves 57 and 58 in the graph of Figure 7 show the results of numerical simulations of the total power of the FWM signals in a two-stage bias-free device.

To simulate curve 57, the previously described model and corresponding numerical parameters of the optical fibers 36 and 38 of both device stages were used.

Curve 58, on the other hand, results from a numerical simulation in which the same values of all other parameters were used and the following wavelengths with zero dispersion optical fiber 36 of the first stage were considered:

01 = 1543.1 nm O 2 = 1543.6 nm a 03 = 1543.7 nm.

These values differ slightly from the dispersion values used in the first numerical simulation. The differences between curves 57 and 58 indicate a high sensitivity of interference between the FWM signals generated in two stages to relatively small variations in the zero dispersion wavelength values along the optical fibers. This high sensitivity, together with the corresponding inaccuracy in determining the dispersion properties of the optical fibers, can explain, in the applicant's view, the incomplete agreement between the experimental measurements (especially measurements 56) and the results of the numerical simulation (curve 58).

The optical distortion suppression circuit of the present invention is particularly suitable for use along a multistage optical communication line. It is particularly suitable for use in combination with an amplifier envisaged as a line amplifier, e.g. together with the line amplifier described in Fig. 2.

One possible arrangement is shown in FIG. 1, where the optical distortion suppression circuit is located at the output of the amplifier in the direction of propagation of the optical signals.

A double pumping line amplification stage comprising a circuit for suppressing optical distortion caused by a four-wave interaction is shown in Fig. 8. The components correspond to those in Fig. 2 already described and are designated with the same reference numbers.

Such an amplifier comprises a single erbium-doped active fiber 62 and a corresponding pumping laser 64 coupled via a dichroic coupler 63; the first optical insulator 61 is positioned toward the active fiber 62 in the signal propagation direction being amplified, while one optical distortion suppression circuit 10, already described in Figure 3, is located away from the active fiber 62.

The optical circulator in the optical circuit 10 prevents the propagation of backscattered signals or optical distortion of any kind in the opposite direction to the communication signals.

The amplifier further comprises a second erbium-doped active fiber 66 coupled to a corresponding pumping laser 68 via a dichroic coupler 67; the optical insulator is positioned away from the fiber 66.

Advantageously, the properties and types of components corresponding to those in the two-stage amplifier of FIG. 2 can be selected in the same variants as in this case.

The active fiber lengths 62 and 66 of both stages and the power emitted by the pumping lasers 64 and 68 are preferably selected such that the total input power (signals plus spontaneous emission) is about -16 dBm, the total power entering the second stage is about 7 dBm and the total output power at the second stage output about 13 dBm.

Placing the optical circuit 10 between the two stages of the amplifier allows to improve the noise figure and minimize the increase in loss on the link to the optical circuit; in fact, the use of the second saturation stage largely compensates for the attenuation in circuit 10, and the drop in the total optical output power of the amplifier caused by the optical circuit is limited to about 1 dB.

If the optical circuit 10 is connected to the nobo from the line amplifier, it is possible to compensate for the attenuation caused by the serial connection of the optical circuit by a short portion of active optical fiber fed via a dichroic coupler with low power pumping radiation (supplied by a small semiconductor laser) amplified by a value corresponding to the attenuation in the optical circuit 10.

Irrespective of the location of the optical circuit 10 relative to the line amplifier, said circuit preferably filters out the spontaneous emission propagating along the optical communication line and which can be generated by the amplifier itself. The amplifier associated with the distortion suppression circuit actually amplifies the communication signals and attenuates radiation at other wavelengths outside the filter reflection bands in the circuit 10.

As an alternative to the description of FIG. 8, the line amplifier may also be configured in a single-stage pumping configuration based on particular application requirements. Also in this case, the FWM optical distortion suppression circuit can advantageously be coupled to an amplifier. In the configuration shown in FIG. 9, a circuit 10 of the same type as described in FIG. 3 is coupled along the amplifier optical fiber 114. The communication signals transmitted by the optical insulator 111 are summed by the dichroic coupler 112 with pumping radiation from the source 113. In the intermediate position along the active fiber 114, a suitable tap for pumping the dichroic couplers 115 and 116 is formed so that radiation in the signal wavelength passes through the optical circuit 10. A second optical insulator 117 is disposed at the end of the active fiber

114.

The optical distortion suppression circuit 10 described in FIG. 3 makes it possible to decorate the FWM signals generated at various stages of the optical communication line by intermodulating between communication signals propagating in a given direction along the line itself.

Fig. 10 shows an optical distortion suppression circuit 10 'in a bidirectional optical WDM communication system according to an alternative version of the present invention.

The circuit 10 'is connected along a bidirectional optical communication line comprising passive optical fiber sections alternating with bidirectional optical amplifiers.

The circuit shown in FIG. 10 is designed for four communication channels in each direction. The shown device can be set up in the same way as the unidirectional circuit in Fig. 3 depending on the number of channels actually used for transmission in each direction.

The optical circuit 10 'consists of an optical circulator 15' with four access gates in the order 11, 12, 13, 14.

The circuit 10 'will now be shown in the case of four optical signals coming from the communication line section connected to the optical circulator gate 11 with wavelengths λ, λ, λ, Xg, Xg <and four additional optical signals coming from the communication fiber section connected to gate 13 of an optical circulator with wavelengths λη, λθ, λ ₈ , λ ₈ . Corresponding line widths are denoted Δνχ, Avg, Avg, ν ₂ , ν ₇ , Δν ₈ , 8ν ₈ , Δν ₈

A selective reflection optical filter 16 at wavelength λ _{7 is} connected to the gate 12 of the optical circulator _{15 '} . A selective delay wavelength 17 is connected to the optical filter 16 and comprises a section of a single-mode optical fiber 18 which is connected to the output of the filter 16 and the other end is connected to the selective reflection optical filter 19 at the wavelength Xg. Thus, the single-mode optical fiber section 18 is longer or longer than half the maximum value of I _c for coherent lengths ν / πΔν], ν / πΔν2, v / mAvg, ν / πΔν2, ν / πΔν ₇ , ν / πΔν ₈ , ν / πΔν ₈ , ν / πΔν ₈ sources 1, 2, 2 ', 2 /', 7, 8_, 8 /, 8 / 'in the optical fiber, where v is the speed of optical radiation propagation in the fiber.

An optical filter 76 with selective reflection at wavelength λ ₇ is connected to the gate 14 of the optical circulator 15 '. Connected to the optical filter 76 is a wavelength selective delay circuit 77 comprising a single-mode optical fiber section 78, one end of which is connected to a selective reflection optical filter at a wavelength λθ. Section 78 of single-mode optical fiber is as long as I _c / 2 or longer.

The optical circuit 10 'consists of other selective delay circuits 17', 17 '', 77 ', 77'', each containing a section 18', 18 '', 7 8 ', 7 8 of an optical single-mode fiber as I _c / 2 or longer with one end connected to the filter output of the preceding selective delay circuit and the other end connected to the optical filter input 19 ', 19'',Tj', 79 '' with selective reflection at a wavelength y> ^ 2 / λθ, λθ

The gates 11 and 13 of the optical circulator 15 are for connection along a communication line, preferably near bidirectional optical amplifiers.

For signals at wavelengths λη, λθ, λθ, λθ · coming from the communication line section connected to the optical circulator gate 11, the function of the circuit is the same as the circuit 10 in Fig. 3 with respect to the foregoing description.

The path of the signals at the wavelengths λη, λθ, λθ, λθ 'coming from the communication line section connected to the optical circulator gate 13, extends from the optical circulator gate 14, comes back to the same gate after reflection from the filter 76 or selective delay circuits 77, 77 77 '' and extends from gate 11 of the optical circulator. Conversely, the wavelength signals outside the bands reflected by the filters 76, 79, 79 ', 7 ', 7 ' ' come from the optical circuit 10 ' through the terminal 24.

This circuit reduces interference between FWM signals generated in both directions along the communication line.

WCC -46

Claims (18)

An optical telecommunications system comprising: at least two optical signal sources modulated at different wavelengths with corresponding coherence times; a multiplexer for multiplexing said signals into a single common optical fiber; - an optical fiber line connected at one end to said common optical fiber of said multiplexer; means for receiving said signals connected to the other end of said optical fiber line and comprising a demultiplexer of said optical signals; characterized in that it comprises an element for suppressing the four-wave interaction between said signals, said element being optically connected in series along said optical fiber line and comprising an optical circuit comprising at least two wave-selective filters, each with a band comprising one of said optical signals, said filters The length of the at least one portion of said optical path comprised between the two successive filters is longer than the length corresponding to the coherence time of at least one of said optical signal sources. An optical telecommunication system according to claim 1, wherein said optical circuit for each said signal comprises a wave-selective waveband filter comprising a corresponding optical signal and not including the remaining optical signals, wherein said filters are optically connected to each other serially in an optical path, wherein the lengths of sections of said optical path between two successive filters are greater than a length corresponding to coherence times of each of said optical signal sources. 3. The optical telecommunication system of claim 1, wherein said optical circuit comprises an optical circulator having an input gate and an output gate connected to said fiber optic line, and at least one input / output gate connected to one of said wavelengths. of selective filters. 4. The optical telecommunication system of claim 1 wherein said wave selective filters are Bragg lattice filters. 5. The optical telecommunication system of claim 4 wherein said Bragg grating filters are made of optical fiber. 6. The optical telecommunication system of claim 1 wherein the at least one optical amplifier is disposed along the optical fiber. An optical telecommunication system according to claim 6, characterized in that said optical amplifier consists of one active optical fiber doped with a fluorescent dopant and a source of pumping radiation. 8. The optical telecommunication system of claim 7, wherein said fluorescent dopant is erbium. 9. The optical telecommunication system of claim 7 wherein the FWM (four-wave interaction) anti-distortion element is optically coupled in series along said active optical fiber at an intermediate position. An optical telecommunication system according to claim 9, characterized in that the outer optical path of the pumping radiation outside said FMW suppression element is formed by two sections into which the active optical fiber is divided by said element. The optical telecommunication system of claim 9, wherein said optical amplifier comprises two sections of active optical fiber, each comprising a source of pumping radiation. 12. An optical telecommunication system according to claim 9, wherein the length of said active optical fiber, the fluorescent dopant concentration and the pumping source power are operatively selected in such a way that the total gain of said amplifier differs by less than 2 dB from that of the same amplifier. said FWM suppression element such that optical continuity between the two active optical fiber sections is maintained. A process for transmitting optical signals comprising: generating two modulated optical signals with corresponding wavelengths; multiplexing said signals at one end of an optical transmission line comprising at least one single-mode optical fiber section in which intermodulation is generated by a four-wave interaction between the signals; - receiving said signals at the other end of said optical transmission line; characterized in that it comprises a selective delay function of said signals at an intermediate position along said optical transmission line, wherein said selective delay function comprises: selectively transmitting said optical signals to corresponding optical paths of a predetermined length such that the lengths are adapted to phase decorrelate said signals to each other; reassembling said signals after passing optical paths; - in which at least common. one section said optical paths is 14. Transmission process optical signals according to claim 13 characterized by that it contains optical amplification step 57 said signals at least once along said optical transmission line. 15. The optical signal transmission process of claim 14, wherein said step of selectively delaying said signals immediately precedes or follows the steps of optically amplifying said signals. 16. An optical amplifier comprising: a first and a second active optical fiber doped with a fluorescent dopant; pumping means for said first and second active optical fibers adapted to deliver pumping energy, coupling means between said first active optical fiber for coupling said optical pumping power and at least two transmission signals at different wavelengths with corresponding coherence times; an element for suppressing the four-wave interaction between said signals, the element being optically connected in series between the first and second active optical fibers; characterized in that said FWM suppression element comprises an optical circuit comprising optical paths of different lengths through which said signals are selectively transmitted, and said lengths being such that at least two signals pass a delay greater than the corresponding coherence times. 17. The optical amplifier of claim 16 wherein at least one portion of said optical paths is common to said signals. 18. The optical amplifier of claim 16, wherein said fluorescent dopant is erbium. 19. The optical amplifier of claim 18 wherein said active fiber comprises aluminum, germanium, and lanthanum as additional dopants.